Thanks for keeping us all informed of the latest technologies and product offerings. Your channel is my favourite go-to resource for glucose monitoring.
Thank Emil. A lot of BS from Roche: They are behind dexcom on CGM field and need data. that's why they need internet access to collect your data. The training of a model may require some compute power, but once a model is trained, then the inference can be done very fast on a mobile cpu. I want my medical data local, combine it with other data coming from my watch (activity), and other sensors, maybe even my calendar and location. And I want to control my own software to integrate it all. Please roche... just make the devices and provide your users with the API's to connect to your devices and diabetics will make the best of your devices by writing wonderful software.
Thanks, Kamil, I’ve been really curious about this one. This was a great explanation of the difference. This cgm is similar to the algorithms we had in DIY and #WeAreNotWaiting
I don't understand how a 2 hour prediction can possibly be relied on. All kinds of things affect your blood sugar. How is it going to take into account that I always rise when I'm driving because of anxiety? How is it going to predict all these other unseen factors that can cause wild swings in blood sugar? I like my dexcom just fine thanks. I don't want to put my life into AI hands (at least not yet, the tech is too new and untested). It's certainly an interesting device but I'll wait and see how this goes. Also having to be online with it for the AI to work? No thank you. ✌
Can I suggest a cheap wireless mic system for your otherwise great videos? Would help out a lot for the audio quality. Hollyland Lark M2 Røde Wireless Go II DJI Mic 2
Appreciate this video. We need to draw a strong distinction between a 'smart algorithm' and actual AI. The algorithm in my tandem t:slim is phenomenal and I'm so glad I made the switch to their control-IQ technology. However, I would encourage anyone to research AI capabilities before giving them any medical decision-making authority. In its present state, AI can only reasonably used with a human chaperone to alert it when decisions are bad. This is fine for training AI, and it is alright when there exists a human chaperone, but for obvious reason it needs to be kept far away from anything it can seriously damage, especially in an insulin pump with no oversight.
Good observation, I don’t really see any critical journalism but rather those companies are having spotlights to showcase their product. It really does feel like blunt advertisement.
I think calling the channel a PR account for big pharma is a little harsh an not helpful critisim at all. I much rather have someone talking and comparing all or most that is out there than having to rely on the websites of the companies only. It is a very informativ video! What it maybe could have more of is talking about the cons. But without using it it could also be a little unfair. I think he is pretty objectiv and just tells what he knows. I honestly can understand his excitedment for anything new. Especially when one has diabetes for longer.
Have you ever heard of "citizen research"? Where does Camile claim to be a critical journalist? He is a diabetic himself, as we all know and from that user based experience he investigates all there is on te market for diabetes management. Your reply is ignorant.
Thanks for keeping us all informed of the latest technologies and product offerings. Your channel is my favourite go-to resource for glucose monitoring.
Thank Emil.
A lot of BS from Roche: They are behind dexcom on CGM field and need data. that's why they need internet access to collect your data. The training of a model may require some compute power, but once a model is trained, then the inference can be done very fast on a mobile cpu. I want my medical data local, combine it with other data coming from my watch (activity), and other sensors, maybe even my calendar and location. And I want to control my own software to integrate it all.
Please roche... just make the devices and provide your users with the API's to connect to your devices and diabetics will make the best of your devices by writing wonderful software.
Thanks, Kamil, I’ve been really curious about this one. This was a great explanation of the difference. This cgm is similar to the algorithms we had in DIY and #WeAreNotWaiting
I don't understand how a 2 hour prediction can possibly be relied on. All kinds of things affect your blood sugar. How is it going to take into account that I always rise when I'm driving because of anxiety? How is it going to predict all these other unseen factors that can cause wild swings in blood sugar?
I like my dexcom just fine thanks. I don't want to put my life into AI hands (at least not yet, the tech is too new and untested). It's certainly an interesting device but I'll wait and see how this goes. Also having to be online with it for the AI to work? No thank you. ✌
AI is a learning program. So all factors you mention should be inputed by you as events in order for AI to learn these and then to predict.
Can I suggest a cheap wireless mic system for your otherwise great videos? Would help out a lot for the audio quality.
Hollyland Lark M2
Røde Wireless Go II
DJI Mic 2
Appreciate this video. We need to draw a strong distinction between a 'smart algorithm' and actual AI. The algorithm in my tandem t:slim is phenomenal and I'm so glad I made the switch to their control-IQ technology. However, I would encourage anyone to research AI capabilities before giving them any medical decision-making authority. In its present state, AI can only reasonably used with a human chaperone to alert it when decisions are bad. This is fine for training AI, and it is alright when there exists a human chaperone, but for obvious reason it needs to be kept far away from anything it can seriously damage, especially in an insulin pump with no oversight.
Except there is no hard line and the exact same thing, it's only algorithms of different complexity. AI is only a salesword today.
This channel REEKS of access " journalism," it's basically a PR account for big pharma .
That comment is much more accurate than roche cgm ^^
Good observation, I don’t really see any critical journalism but rather those companies are having spotlights to showcase their product. It really does feel like blunt advertisement.
I think calling the channel a PR account for big pharma is a little harsh an not helpful critisim at all. I much rather have someone talking and comparing all or most that is out there than having to rely on the websites of the companies only. It is a very informativ video! What it maybe could have more of is talking about the cons. But without using it it could also be a little unfair. I think he is pretty objectiv and just tells what he knows. I honestly can understand his excitedment for anything new. Especially when one has diabetes for longer.
Have you ever heard of "citizen research"? Where does Camile claim to be a critical journalist? He is a diabetic himself, as we all know and from that user based experience he investigates all there is on te market for diabetes management. Your reply is ignorant.
Love your videos, thank you for the news
Hello, I'm not interested by this cgm. I don't get what's good about it.
If it's not iCGM it's pretty much a no-go for me. I also call BS on it needing an internet connection to give AI predictions.
Why should it require calibration even when it has AI?
I don't see any advantage to using this. I am happy with my G7